Undecidability of the block gluing classes of homshifts

TL;DR

This paper proves that it is undecidable to determine certain mixing properties of homshifts, a class of multidimensional shifts of finite type, by linking the problem to the finiteness of finitely presented groups.

Contribution

It demonstrates the undecidability of the block gluing property in homshifts, extending the scope of undecidability results in symbolic dynamics using algebraic topology.

Findings

Undecidability of whether a homshift is $ heta(n)$-block gluing.

Finiteness of finitely presented groups relates to homshift properties.

Topological interpretation of the square cover aids in proving undecidability.

Abstract

A homshift is a $d$-dimensional shift of finite type which arises as the space of graph homomorphisms from the grid graph $\mathbb Z^d$ to a finite connected undirected graph $G$. While shifts of finite type are known to be mired by the swamp of undecidability, homshifts seem to be better behaved and there was hope that all the properties of homshifts are decidable. In this paper we build on the work by Gangloff, Hellouin de Menibus and Oprocha (arxiv:2211.04075) to show that finer mixing properties are undecidable for reasons completely different than the ones used to prove undecidability for general multidimensional shifts of finite type. Inspired by the work of Gao, Jackson, Krohne and Seward (arxiv:1803.03872) and elementary algebraic topology, we interpret the square cover introduced by Gangloff, Hellouin de Menibus and Oprocha topologically. Using this interpretation, we prove that it is undecidable whether a homshift is $\Theta(n)$-block gluing or not, by relating this problem to the one of finiteness for finitely presented groups.